3.2.1 Data

viz_timelineW <- dat_orderlines %>% 
    mutate(sale_week = FLOOR_WEEK(sale_date)) %>% 
    group_by(sale_week) %>% 
    summarise(total_spend = sum(price)) %>% 
    ggplot(aes(x = sale_week, 
        y = total_spend, 
        fill = total_spend,
        text = str_glue("Date: {sale_week}
                         Revenue: {scales::dollar(total_spend)}"))) +
    geom_smooth() +
    geom_col() +
    theme_ipsum() +
    labs(
        title = "Sales over Time",
        subtitle = "",
        caption = "",
        x = "Week",
        y = "Total Spend",
        fill = "Total Spend"#,
       # color = "Persona"
       ) +
    viridis::scale_fill_viridis(direction = 1, discrete = FALSE) +
    theme(
      legend.position = "bottom",
#     axis.text = element_blank(),
#     axis.text.x = element_text(angle = 90),
#     axis.title = element_blank(),
#      panel.grid = element_blank(),
     plot.margin = margin(.5, .5, .5, .5, "cm")
) 

3.2.3 Static

viz_timelineW

3.3.1 Data

stat_items <- dat_orderlines %>% 
    group_by(item_group, item_name) %>% 
    summarise(n_items = n(),
              n_orders = n_distinct(order_id),
              sum_price = sum(price)
              ) %>% ungroup() %>% 
    mutate(
        revenue = sum_price %>% scales::dollar(accuracy = 1,big.mark = ","),
        perc_total = sum_price/sum(sum_price),
        percent_total = perc_total %>% scales::percent(accuracy = 0.01),
        rankv = -perc_total
    ) %>% select(item_group, item_name, 
                 n_items, n_orders, revenue, percent_total, 
                 everything()) %>% 
    arrange(desc(sum_price),n_orders, n_items) %>% 
    mutate(rank = rank(rankv))

## `summarise()` has grouped output by 'item_group'. You can override using the
## `.groups` argument.

viz_items <- stat_items %>% 
    filter(perc_total >= 0.01) %>% 
    ggplot(aes(x = item_name %>% fct_reorder(sum_price), 
               y = sum_price, 
               fill = sum_price,
               text = str_glue(" 
                               Name: {item_name}
                               Revenue: {revenue}
                               % Revenue: {percent_total}
                               Items Sold: {n_items}
                               Total Orders: {n_orders}"))) +
    geom_col()+
    theme_ipsum() +
  #  ylim(0, 750) +
    coord_flip()  +
    labs(
        title = "Top Performing Products",
        subtitle = "Total Spend vs Product",
        caption = "",
        x = "Product",
        y = "Total Spend",
        fill = "Total Spend") +
    viridis::scale_fill_viridis(direction = -1, 
                                discrete = FALSE, 
                                option = "plasma") +
    theme(
      legend.position = "bottom",
      panel.grid = element_blank(),
      plot.margin = margin(.5, .5, .5, .5, "in")
) 

3.3.3 Static

viz_items

3.4.1 Data

stat_groups <- dat_orderlines %>% 
    filter(item_group != "Heat Pack") %>% 
    group_by(item_group) %>% 
    mutate(receipt = paste0(item_name)) %>% 
    summarise(items_sold = toString(unique(receipt)),
              n_items = n(),
              n_orders = n_distinct(order_id),
              sum_price = sum(price)
              ) %>% ungroup() %>% 
    mutate(
        revenue = sum_price %>% scales::dollar(accuracy = 1),
        perc_total = sum_price/sum(sum_price),
        percent_total = perc_total %>% scales::percent(accuracy = 0.01),
        rankv = -perc_total
    ) %>% select(item_group, items_sold, 
                 n_items, n_orders, revenue, percent_total,
                 everything()) %>% 
    arrange(desc(sum_price),n_orders, item_group) %>% 
    mutate(rank = rank(rankv))

viz_groups <- stat_groups %>% 
    filter(rank <= 10) %>% 
    ggplot(aes(x = item_group %>% fct_reorder(sum_price), 
               y = sum_price, 
               fill = sum_price,
               text = str_glue(" 
                               Group: {item_group}
                               Revenue: {revenue}
                               % Revenue: {percent_total}
                               Products Sold: {n_items}
                               Total Orders: {n_orders}
                               Products Sold: {items_sold}"
               ))) +
    geom_col()+
    coord_flip() +
    theme_ipsum()  +
    labs(
        title = "Top Performing Product Groups",
        subtitle = "Total Spend vs Product Groups",
        caption = "",
        x = "Product Group",
        y = "Total Spend",
        fill = "Total Spend") +
    theme(
      legend.position = "bottom",
      plot.margin = margin(.5, .5, .5, .5, "cm")) +
    viridis::scale_fill_viridis(direction = -1, 
                                discrete = FALSE, 
                                option = "plasma") +
    theme(
      legend.position = "bottom",
      panel.grid = element_blank(),
      plot.margin = margin(.5, .5, .5, .5, "in")
) 

3.4.3 Static

viz_groups

4.2.1 Recency

 viz_recency <- dat_customers %>% 
    ggplot(aes(x = rfm_persona_rev, 
               y = recency, 
               fill = rfm_persona_rev
               )) +
    ggdist::stat_halfeye(aes(color = rfm_persona_rev),
        adjust = 0.5, 
        justification = -.2, 
        .width = 0,
        point_colour = NA, 
        na.rm = T) +
    geom_boxplot(aes(color = rfm_persona_rev),
        width = .4, 
        outlier.color = NA, 
        outlier.alpha =  0.33,
        alpha = 0.66, 
        na.rm = T) +
    ggdist::stat_dots(aes(color = rfm_persona_rev),
         side = "left", 
         justification = 1.1, 
         binwidth = .5) +
    theme_ipsum() +
    ylim(0, 365) +
    coord_flip()  +
    labs(
        title = "Persona Distributions: Recency",
        subtitle = "Recency Raincloud",
        caption = "Here we can see the persona definitions more clearly as the different groups are heavily driven by recency.",
        x = "Persona",
        y = "Recency",
        fill = "Persona",
        color = "Persona") +
    viridis::scale_fill_viridis(direction = -1, 
                                discrete = TRUE, 
                                option = "plasma") +
    viridis::scale_color_viridis(direction = -1, 
                                discrete = TRUE, 
                                option = "plasma") +
    theme(
      legend.position = "bottom",
      panel.grid = element_blank(),
      plot.margin = margin(.5, .5, .5, .5, "in")
) 

viz_recency

4.2.2 Frequency

 viz_frequency <- dat_customers %>% 
    ggplot(aes(x = rfm_persona_rev, 
               y = frequency, 
               fill = rfm_persona_rev
               )) +
    ggdist::stat_halfeye(aes(color = rfm_persona_rev),
        adjust = 0.5, 
        justification = -.2, 
        .width = 0,
        point_colour = NA, 
        na.rm = T) +
    geom_boxplot(aes(color = rfm_persona_rev),
        width = .4, 
        outlier.color = NA, 
        outlier.alpha =  0.33,
        alpha = 0.66, 
        na.rm = T) +
    # ggdist::stat_dots(aes(color = rfm_persona_rev),
    #      side = "left", 
    #      justification = 1.1, 
    #      binwidth = .01) +
    theme_ipsum() +
    ylim(1,5) +
    coord_flip()  +
    labs(
        title = "Persona Distributions: Frequency",
        subtitle = "Frequency Raincloud",
        caption = "Here we see CHAMPIONS and CAN'T LOSE reaching into the higher frequency tiers via repeated purchases",
        x = "Persona",
        y = "Monetary",
        fill = "Persona",
        color = "Persona") +
    viridis::scale_fill_viridis(direction = -1, 
                                discrete = TRUE, 
                                option = "plasma") +
    viridis::scale_color_viridis(direction = -1, 
                                discrete = TRUE, 
                                option = "plasma") +
    theme(
      legend.position = "bottom",
      panel.grid = element_blank(),
      plot.margin = margin(.5, .5, .5, .5, "in")
) 

viz_frequency